The overall objective of this research is to make room-temperature semiconductor gamma-ray detector arrays into practical tools for clinical nuclear medicine, especially in the diagnosis and management of cancer. Previous work has demonstrated that semiconductor detector arrays offer unprecedented spatial and energy resolution. The applicants have designed, built and successfully tested several high-resolution detector arrays based on readout integrated circuits called multiplexers, and have shown that use of small pixels leads to dramatic improvements in the energy spectra of these devices. The research proposed here will further develop the scientific and engineering foundations of semiconductor-array technology. These models will then be used to develop optimal estimators of the interaction coordinates and gamma-ray energy. The Fisher information matrix will be used to assess the accuracy of the estimates and how it depends on system geometry and material properties. Then the estimation methods will be extended to determination of the spatial and spectral distribution of gamma-ray fluence on the detector. The newly developed concepts of list-mode likelihood and list-mode Fisher information will be used here. The knowledge gained will be used to design an improve "smart" multiplexer readout. A parallel effort will continue to investigate properties of various semiconductor materials, including CdZnTe, CdTe, T1NBr, HgI2 and PbI2, with the objective of producing a particular detector with spatial and energy resolution, space-bandwidth product and count-rate capability far in excess of what is available today. On the technological front, the applicants propose several ways to reduce the cost of these devices, improve their yield and reliability, and produce imaging modules that can serve as building blocks for sophisticated SPECT systems. Finally, they propose to continue on-going investigation of novel imaging configurations to take maximum advantage of the unique new capabilities offered by semiconductor detector arrays.